sqlmesh 0.209.1.dev66__tar.gz → 0.209.1.dev67__tar.gz

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sqlmesh
-Version: 0.209.1.dev66
+Version: 0.209.1.dev67
 Summary: Next-generation data transformation framework
 Author-email: "TobikoData Inc." <engineering@tobikodata.com>
 License:                                  Apache License

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/_version.py

@@ -28,7 +28,7 @@ version_tuple: VERSION_TUPLE
 commit_id: COMMIT_ID
 __commit_id__: COMMIT_ID
 
-__version__ = version = '0.209.1.dev66'
-__version_tuple__ = version_tuple = (0, 209, 1, 'dev66')
+__version__ = version = '0.209.1.dev67'
+__version_tuple__ = version_tuple = (0, 209, 1, 'dev67')
 
-__commit_id__ = commit_id = 'g1e2760f97'
+__commit_id__ = commit_id = 'gf73cdfe49'

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/engine_adapter/athena.py

@@ -45,6 +45,7 @@ class AthenaEngineAdapter(PandasNativeFetchDFSupportMixin, RowDiffMixin):
     # >>> self._execute('/* test */ DESCRIBE foo')
     #     pyathena.error.OperationalError: FAILED: ParseException line 1:0 cannot recognize input near '/' '*' 'test'
     ATTACH_CORRELATION_ID = False
+    SUPPORTED_DROP_CASCADE_OBJECT_KINDS = ["DATABASE", "SCHEMA"]
 
     def __init__(
         self, *args: t.Any, s3_warehouse_location: t.Optional[str] = None, **kwargs: t.Any
@@ -314,13 +315,13 @@ class AthenaEngineAdapter(PandasNativeFetchDFSupportMixin, RowDiffMixin):
 
         return None
 
-    def drop_table(self, table_name: TableName, exists: bool = True) -> None:
+    def drop_table(self, table_name: TableName, exists: bool = True, **kwargs: t.Any) -> None:
         table = exp.to_table(table_name)
 
         if self._query_table_type(table) == "hive":
             self._truncate_table(table)
 
-        return super().drop_table(table_name=table, exists=exists)
+        return super().drop_table(table_name=table, exists=exists, **kwargs)
 
     def _truncate_table(self, table_name: TableName) -> None:
         table = exp.to_table(table_name)

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/engine_adapter/base.py

@@ -108,6 +108,7 @@ class EngineAdapter:
     SUPPORTS_CLONING = False
     SUPPORTS_MANAGED_MODELS = False
     SUPPORTS_CREATE_DROP_CATALOG = False
+    SUPPORTED_DROP_CASCADE_OBJECT_KINDS: t.List[str] = []
     SCHEMA_DIFFER = SchemaDiffer()
     SUPPORTS_TUPLE_IN = True
     HAS_VIEW_BINDING = False
@@ -1044,14 +1045,14 @@ class EngineAdapter:
                 f"Can't drop data object '{data_object.to_table().sql(dialect=self.dialect)}' of type '{data_object.type.value}'"
             )
 
-    def drop_table(self, table_name: TableName, exists: bool = True) -> None:
+    def drop_table(self, table_name: TableName, exists: bool = True, **kwargs: t.Any) -> None:
         """Drops a table.
 
         Args:
             table_name: The name of the table to drop.
             exists: If exists, defaults to True.
         """
-        self._drop_object(name=table_name, exists=exists)
+        self._drop_object(name=table_name, exists=exists, **kwargs)
 
     def drop_managed_table(self, table_name: TableName, exists: bool = True) -> None:
         """Drops a managed table.
@@ -1067,6 +1068,7 @@ class EngineAdapter:
         name: TableName | SchemaName,
         exists: bool = True,
         kind: str = "TABLE",
+        cascade: bool = False,
         **drop_args: t.Any,
     ) -> None:
         """Drops an object.
@@ -1077,8 +1079,13 @@ class EngineAdapter:
             name: The name of the table to drop.
             exists: If exists, defaults to True.
             kind: What kind of object to drop. Defaults to TABLE
+            cascade: Whether or not to DROP ... CASCADE.
+                Note that this is ignored for :kind's that are not present in self.SUPPORTED_DROP_CASCADE_OBJECT_KINDS
             **drop_args: Any extra arguments to set on the Drop expression
         """
+        if cascade and kind.upper() in self.SUPPORTED_DROP_CASCADE_OBJECT_KINDS:
+            drop_args["cascade"] = cascade
+
         self.execute(exp.Drop(this=exp.to_table(name), kind=kind, exists=exists, **drop_args))
 
     def get_alter_operations(

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/engine_adapter/base_postgres.py

@@ -24,6 +24,7 @@ class BasePostgresEngineAdapter(EngineAdapter):
     DEFAULT_BATCH_SIZE = 400
     COMMENT_CREATION_TABLE = CommentCreationTable.COMMENT_COMMAND_ONLY
     COMMENT_CREATION_VIEW = CommentCreationView.COMMENT_COMMAND_ONLY
+    SUPPORTED_DROP_CASCADE_OBJECT_KINDS = ["SCHEMA", "TABLE", "VIEW"]
 
     def columns(
         self, table_name: TableName, include_pseudo_columns: bool = False

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/engine_adapter/bigquery.py

@@ -66,6 +66,7 @@ class BigQueryEngineAdapter(InsertOverwriteWithMergeMixin, ClusteredByMixin, Row
     SUPPORTS_CLONING = True
     MAX_TABLE_COMMENT_LENGTH = 1024
     MAX_COLUMN_COMMENT_LENGTH = 1024
+    SUPPORTED_DROP_CASCADE_OBJECT_KINDS = ["SCHEMA"]
 
     SCHEMA_DIFFER = SchemaDiffer(
         compatible_types={

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/engine_adapter/clickhouse.py

@@ -616,6 +616,7 @@ class ClickhouseEngineAdapter(EngineAdapterWithIndexSupport, LogicalMergeMixin):
         name: TableName | SchemaName,
         exists: bool = True,
         kind: str = "TABLE",
+        cascade: bool = False,
         **drop_args: t.Any,
     ) -> None:
         """Drops an object.
@@ -628,7 +629,6 @@ class ClickhouseEngineAdapter(EngineAdapterWithIndexSupport, LogicalMergeMixin):
             kind: What kind of object to drop. Defaults to TABLE
             **drop_args: Any extra arguments to set on the Drop expression
         """
-        drop_args.pop("cascade", None)
         self.execute(
             exp.Drop(
                 this=exp.to_table(name),

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/engine_adapter/duckdb.py

@@ -37,6 +37,7 @@ class DuckDBEngineAdapter(LogicalMergeMixin, GetCurrentCatalogFromFunctionMixin,
     COMMENT_CREATION_TABLE = CommentCreationTable.COMMENT_COMMAND_ONLY
     COMMENT_CREATION_VIEW = CommentCreationView.COMMENT_COMMAND_ONLY
     SUPPORTS_CREATE_DROP_CATALOG = True
+    SUPPORTED_DROP_CASCADE_OBJECT_KINDS = ["SCHEMA", "TABLE", "VIEW"]
 
     @property
     def catalog_support(self) -> CatalogSupport:

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/engine_adapter/snowflake.py

@@ -55,6 +55,7 @@ class SnowflakeEngineAdapter(GetCurrentCatalogFromFunctionMixin, ClusteredByMixi
     SUPPORTS_MANAGED_MODELS = True
     CURRENT_CATALOG_EXPRESSION = exp.func("current_database")
     SUPPORTS_CREATE_DROP_CATALOG = True
+    SUPPORTED_DROP_CASCADE_OBJECT_KINDS = ["DATABASE", "SCHEMA", "TABLE"]
     SCHEMA_DIFFER = SchemaDiffer(
         parameterized_type_defaults={
             exp.DataType.build("BINARY", dialect=DIALECT).this: [(8388608,)],

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/engine_adapter/spark.py

@@ -57,6 +57,7 @@ class SparkEngineAdapter(
     # currently check for storage formats we say we don't support REPLACE TABLE
     SUPPORTS_REPLACE_TABLE = False
     QUOTE_IDENTIFIERS_IN_VIEWS = False
+    SUPPORTED_DROP_CASCADE_OBJECT_KINDS = ["DATABASE", "SCHEMA"]
 
     WAP_PREFIX = "wap_"
     BRANCH_PREFIX = "branch_"

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/engine_adapter/trino.py

@@ -53,6 +53,7 @@ class TrinoEngineAdapter(
     COMMENT_CREATION_TABLE = CommentCreationTable.IN_SCHEMA_DEF_NO_CTAS
     COMMENT_CREATION_VIEW = CommentCreationView.COMMENT_COMMAND_ONLY
     SUPPORTS_REPLACE_TABLE = False
+    SUPPORTED_DROP_CASCADE_OBJECT_KINDS = ["SCHEMA"]
     DEFAULT_CATALOG_TYPE = "hive"
     QUOTE_IDENTIFIERS_IN_VIEWS = False
     SCHEMA_DIFFER = SchemaDiffer(

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh/core/snapshot/evaluator.py

@@ -1241,6 +1241,10 @@ class SnapshotEvaluator:
                     table_name,
                     is_table_deployable=is_table_deployable,
                     physical_schema=snapshot.physical_schema,
+                    # we need to set cascade=true or we will get a 'cant drop because other objects depend on it'-style
+                    # error on engines that enforce referential integrity, such as Postgres
+                    # this situation can happen when a snapshot expires but downstream view snapshots that reference it have not yet expired
+                    cascade=True,
                 )
             except Exception:
                 # Use `get_data_object` to check if the table exists instead of `table_exists` since the former
@@ -1771,7 +1775,7 @@ class MaterializableStrategy(PromotableStrategy, abc.ABC):
 
     def delete(self, name: str, **kwargs: t.Any) -> None:
         _check_table_db_is_physical_schema(name, kwargs["physical_schema"])
-        self.adapter.drop_table(name)
+        self.adapter.drop_table(name, cascade=kwargs.pop("cascade", False))
         logger.info("Dropped table '%s'", name)
 
     def _replace_query_for_model(
@@ -2372,15 +2376,16 @@ class ViewStrategy(PromotableStrategy):
         )
 
     def delete(self, name: str, **kwargs: t.Any) -> None:
+        cascade = kwargs.pop("cascade", False)
         try:
-            self.adapter.drop_view(name)
+            self.adapter.drop_view(name, cascade=cascade)
         except Exception:
             logger.debug(
                 "Failed to drop view '%s'. Trying to drop the materialized view instead",
                 name,
                 exc_info=True,
             )
-            self.adapter.drop_view(name, materialized=True)
+            self.adapter.drop_view(name, materialized=True, cascade=cascade)
         logger.info("Dropped view '%s'", name)
 
     def _is_materialized_view(self, model: Model) -> bool:

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/sqlmesh.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: sqlmesh
-Version: 0.209.1.dev66
+Version: 0.209.1.dev67
 Summary: Next-generation data transformation framework
 Author-email: "TobikoData Inc." <engineering@tobikodata.com>
 License:                                  Apache License

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/tests/core/engine_adapter/integration/__init__.py

@@ -27,7 +27,7 @@ from _pytest.mark import MarkDecorator
 from _pytest.mark.structures import ParameterSet
 
 if t.TYPE_CHECKING:
-    from sqlmesh.core._typing import TableName
+    from sqlmesh.core._typing import TableName, SchemaName
     from sqlmesh.core.engine_adapter._typing import Query
 
 TEST_SCHEMA = "test_schema"
@@ -222,6 +222,13 @@ class TestContext:
             return self._test_type.split("-", maxsplit=1)[1]
         return None
 
+    @property
+    def engine_type(self) -> str:
+        if self.mark.startswith("gcp_postgres"):
+            return "gcp_postgres"
+
+        return self.mark.split("_")[0]
+
     @property
     def columns_to_types(self):
         if self._columns_to_types is None:
@@ -307,7 +314,7 @@ class TestContext:
     def add_test_suffix(self, value: str) -> str:
         return f"{value}_{self.test_id}"
 
-    def get_metadata_results(self, schema: t.Optional[str] = None) -> MetadataResults:
+    def get_metadata_results(self, schema: t.Optional[SchemaName] = None) -> MetadataResults:
         schema = schema if schema else self.schema(TEST_SCHEMA)
         return MetadataResults.from_data_objects(self.engine_adapter.get_data_objects(schema))
 

{sqlmesh-0.209.1.dev66 → sqlmesh-0.209.1.dev67}/tests/core/engine_adapter/integration/test_integration.py

@@ -21,6 +21,7 @@ from sqlmesh.cli.project_init import init_example_project
 from sqlmesh.core.config import load_config_from_paths
 from sqlmesh.core.config.connection import ConnectionConfig
 import sqlmesh.core.dialect as d
+from sqlmesh.core.environment import EnvironmentSuffixTarget
 from sqlmesh.core.dialect import select_from_values
 from sqlmesh.core.model import Model, load_sql_based_model
 from sqlmesh.core.engine_adapter.shared import DataObject, DataObjectType
@@ -2372,11 +2373,7 @@ def test_init_project(ctx: TestContext, tmp_path_factory: pytest.TempPathFactory
             k: [_normalize_snowflake(name) for name in v] for k, v in object_names.items()
         }
 
-    if ctx.mark.startswith("gcp_postgres"):
-        engine_type = "gcp_postgres"
-    else:
-        engine_type = ctx.mark.split("_")[0]
-    init_example_project(tmp_path, engine_type, schema_name=schema_name)
+    init_example_project(tmp_path, ctx.engine_type, schema_name=schema_name)
 
     config = load_config_from_paths(
         Config,
@@ -3596,3 +3593,137 @@ def test_identifier_length_limit(ctx: TestContext):
         match=re.escape(match),
     ):
         adapter.create_table(long_table_name, {"col": exp.DataType.build("int")})
+
+
+@pytest.mark.parametrize(
+    "environment_suffix_target",
+    [
+        EnvironmentSuffixTarget.TABLE,
+        EnvironmentSuffixTarget.SCHEMA,
+        EnvironmentSuffixTarget.CATALOG,
+    ],
+)
+def test_janitor(
+    ctx: TestContext, tmp_path: pathlib.Path, environment_suffix_target: EnvironmentSuffixTarget
+):
+    if (
+        environment_suffix_target == EnvironmentSuffixTarget.CATALOG
+        and not ctx.engine_adapter.SUPPORTS_CREATE_DROP_CATALOG
+    ):
+        pytest.skip("Engine does not support catalog-based virtual environments")
+
+    schema = ctx.schema()  # catalog.schema
+    parsed_schema = d.to_schema(schema)
+
+    init_example_project(tmp_path, ctx.engine_type, schema_name=parsed_schema.db)
+
+    def _set_config(_gateway: str, config: Config) -> None:
+        config.environment_suffix_target = environment_suffix_target
+        config.model_defaults.dialect = ctx.dialect
+
+    sqlmesh = ctx.create_context(path=tmp_path, config_mutator=_set_config)
+
+    sqlmesh.plan(auto_apply=True)
+
+    # create a new model in dev
+    (tmp_path / "models" / "new_model.sql").write_text(f"""
+        MODEL (
+            name {schema}.new_model,
+            kind FULL
+        );
+
+        select * from {schema}.full_model
+    """)
+    sqlmesh.load()
+
+    result = sqlmesh.plan(environment="dev", auto_apply=True)
+    assert result.context_diff.is_new_environment
+    assert len(result.context_diff.new_snapshots) == 1
+    new_model = list(result.context_diff.new_snapshots.values())[0]
+    assert "new_model" in new_model.name.lower()
+
+    # check physical objects
+    snapshot_table_name = exp.to_table(new_model.table_name(), dialect=ctx.dialect)
+    snapshot_schema = snapshot_table_name.db
+
+    prod_schema = normalize_identifiers(d.to_schema(schema), dialect=ctx.dialect)
+    dev_env_schema = prod_schema.copy()
+    if environment_suffix_target == EnvironmentSuffixTarget.CATALOG:
+        dev_env_schema.set("catalog", exp.to_identifier(f"{prod_schema.catalog}__dev"))
+    else:
+        dev_env_schema.set("db", exp.to_identifier(f"{prod_schema.db}__dev"))
+    normalize_identifiers(dev_env_schema, dialect=ctx.dialect)
+
+    md = ctx.get_metadata_results(prod_schema)
+    if environment_suffix_target == EnvironmentSuffixTarget.TABLE:
+        assert sorted([v.lower() for v in md.views]) == [
+            "full_model",
+            "incremental_model",
+            "new_model__dev",
+            "seed_model",
+        ]
+    else:
+        assert sorted([v.lower() for v in md.views]) == [
+            "full_model",
+            "incremental_model",
+            "seed_model",
+        ]
+    assert not md.tables
+    assert not md.managed_tables
+
+    if environment_suffix_target != EnvironmentSuffixTarget.TABLE:
+        # note: this is "catalog__dev.schema" for EnvironmentSuffixTarget.CATALOG and "catalog.schema__dev" for EnvironmentSuffixTarget.SCHEMA
+        md = ctx.get_metadata_results(dev_env_schema)
+        assert [v.lower() for v in md.views] == ["new_model"]
+        assert not md.tables
+        assert not md.managed_tables
+
+    md = ctx.get_metadata_results(snapshot_schema)
+    assert not md.views
+    assert not md.managed_tables
+    assert sorted(t.split("__")[1].lower() for t in md.tables) == [
+        "full_model",
+        "incremental_model",
+        "new_model",
+        "seed_model",
+    ]
+
+    # invalidate dev and run the janitor to clean it up
+    sqlmesh.invalidate_environment("dev")
+    assert sqlmesh.run_janitor(
+        ignore_ttl=True
+    )  # ignore_ttl to delete the new_model snapshot even though it hasnt expired yet
+
+    # there should be no dev environment or dev tables / schemas
+    md = ctx.get_metadata_results(prod_schema)
+    assert sorted([v.lower() for v in md.views]) == [
+        "full_model",
+        "incremental_model",
+        "seed_model",
+    ]
+    assert not md.tables
+    assert not md.managed_tables
+
+    if environment_suffix_target != EnvironmentSuffixTarget.TABLE:
+        if environment_suffix_target == EnvironmentSuffixTarget.SCHEMA:
+            md = ctx.get_metadata_results(dev_env_schema)
+        else:
+            try:
+                md = ctx.get_metadata_results(dev_env_schema)
+            except Exception as e:
+                # Most engines will raise an error when @set_catalog tries to set a catalog that doesnt exist
+                # in this case, we just swallow the error. We know this call already worked before in the earlier checks
+                md = MetadataResults()
+
+        assert not md.views
+        assert not md.tables
+        assert not md.managed_tables
+
+    md = ctx.get_metadata_results(snapshot_schema)
+    assert not md.views
+    assert not md.managed_tables
+    assert sorted(t.split("__")[1].lower() for t in md.tables) == [
+        "full_model",
+        "incremental_model",
+        "seed_model",
+    ]